{"id":"59a89e86-fc02-48b5-8cd9-50cf879aa8e9","arxiv_id":"2603.01575","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A measurement is a PVM if and only if it is completely intersubjective — it and every outcome-coarse-graining guarantee identical outcomes for all observers — and this gap-free property exactly characterizes classical probabilistic theories.","lead":"This paper proves that a measurement in quantum theory is a 'real observable' — a projective measurement — exactly when it and every merging of its outcomes forces two observers to agree on the result. The same condition separates classical from non-classical theories, and gives a quantitative measure of how much observers agree.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The classicality characterization in Theorem 4 is proven only for finite-dimensional systems, but the abstract and introduction claim it for 'any non-classical theory'; this unsupported infinite-dimensional generalization is the main gap between the headline claim and the proof.","rationale":"I read the main theorems in good faith. Proposition 1 and Theorem 3 are solid: the equivalence between intersubjectivity and sharpness is correctly proved, and the reduction of complete intersubjectivity to elementwise-sharp effects in quantum theory correctly yields PVMs. Theorem 4's proof is more delicate. The reader's weakest_assumption focused on imported structural facts and Lemma 7's terse Krein–Millman step; these are genuine exposition gaps, but they are fillable. The linear-independence characterization of simplex cones is standard, sharp effects attaining value 1 follows because otherwise a nonzero scalar multiple of the effect would lie below both the effect and its complement, and the finite decomposition as a sum of indecomposable effects follows from Krein–Millman on a compact base of the cone plus Carathéodory. So I would not reject or downgrade the finite-dimensional theorem on those grounds. The load-bearing concern I identify instead is the scope mismatch: the advertised principle 'non-classicality is exactly the failure of complete intersubjectivity' is stated for all theories, while the proof establishes it only in finite dimensions. Because the abstract and introduction market the infinite-dimensional version, the paper's central claim as presented is stronger than the evidence. This justifies the CONDITIONAL verdict already given, so I recommend no change to the reader's verdict.","tokens_in":19306,"tokens_out":32559,"duration_ms":318866,"concrete_test":"Attempt to extend Lemma 7 to infinite dimensions by dropping the 'n-dimensional' assumption and checking whether the key step '1_S − εΣ_{i=1}^{n+1} a_i = Σ_j b_j with each b_j an indecomposable effect' remains valid. Since Krein–Millman plus Carathéodory gives finite convex combinations only in finite dimensions, either (a) construct an explicit infinite-dimensional non-classical GPT (or infinite-outcome quantum POVM) where an intersubjective measurement fails to be completely intersubjective, or (b) supply a Choquet-integral substitute and prove that complete intersubjective measurements have controlled outcome cardinality. If neither succeeds, the abstract and introduction must be restricted to 'finite-dimensional systems.'","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 4 is explicitly finite-dimensional: 'A finite-dimensional system is classical iff all intersubjective measurements on it are completely intersubjective.' Yet the abstract states, without qualifier, that 'a system is classical iff intersubjectivity is preserved under any coarse-graining,' and the introduction says the gap between intersubjectivity and complete intersubjectivity is 'common to all non-classical theories.' Infinite dimension is not a harmless qualifier here. The proof of Theorem 4 uses finite-dimensionality at structurally essential points: (i) the simplex characterization imported from [30] is stated for finite-dimensional systems; (ii) Lemma 7 bounds the number of outcomes of a completely intersubjective measurement by the dimension n of the system; and (iii) the Krein–Millman step that 1_S − εΣ a_i decomposes as a finite sum Σ b_j of indecomposable effects relies on Carathéodory's theorem for compact convex sets, which is finite-dimensional. In infinite-dimensional cones, not every element is a finite sum of extreme rays, so the construction of an intersubjective-but-not-completely-intersubjective measurement is not established outside the finite-dimensional setting. The other gaps flagged by the reader are real but fillable: sharp effects attaining norm 1 follows from a compactness argument, and the finite sum decomposition follows from Krein–Millman on a base of the cone. The infinite-dimensional overclaim is the one place where the advertised result exceeds what the supplied arguments can support.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes an operational principle, complete intersubjectivity, in the framework of generalized probabilistic theories (GPTs). It defines α-intersubjectivity via the guaranteed agreement of two observers performing the same measurement, proves its equivalence with sharpness (Prop. 1), and gives a quantitative refinement. In quantum theory it characterizes intersubjective POVMs by pairwise trivial support intersections (Thm. 2), provides explicit intersubjective POVMs that are not PVMs, and proves that a POVM is a PVM iff every coarse-graining is intersubjective (Thm. 3). It also derives quantitative formulas for coin-toss, classical, and qubit measurements (Eqs. (4)–(6)), studies the relation between intersubjectivity, elementwise sharpness, extremality, and complete intersubjectivity, and claims that a finite-dimensional GPT is classical iff all intersubjective measurements are completely intersubjective (Thm. 4). Finally, it argues that (completely) intersubjective measurements are sufficient for state tomography and optimal state discrimination.","tokens_in":19511,"tokens_out":30020,"duration_ms":301221,"significance":"If the main theorems hold, this is an elegant and genuinely operational result. Theorem 3 gives a clean, interpretable criterion that bridges the traditional projection-valued formulation of observables with the modern POVM framework, and the quantitative bounds and explicit counterexamples (Examples 1–5) are concrete and useful. The paper makes appropriate use of strong external results (Gudder's sharp-effect theory, the simplex characterization of classicality in [30], and Bauer's maximum principle), and the proofs of Proposition 1 and Theorem 2 are essentially sound. The main advertised claim about classicality, however, is stated more broadly than what is proved: the abstract presents the classicality characterization without a finite-dimensional qualifier, while Theorem 4 is finite-dimensional and the proof uses finite-dimensionality in structurally essential places. Several steps in the End Matter proofs, particularly in Lemmas 6 and 7, are also under-specified. The central ideas are promising and likely repairable, but the manuscript needs a substantial revision before the claims as advertised are fully supported.","major_comments":[{"comment":"The abstract states without qualification that 'a system is classical iff intersubjectivity is preserved under any coarse-graining,' and the introduction says this gap is 'common to all non-classical theories.' Theorem 4, however, is proved only for finite-dimensional systems. The proof uses finite-dimensionality at essential points: the imported simplex characterization from [30] is finite-dimensional, Lemma 7 uses a finite-dimensional Krein–Millman/Carathéodory step and a finite outcome-count bound, and the construction of n+1 indecomposable effects is a finite-dimensional cone argument. No argument is supplied for infinite-dimensional state spaces. Please either prove the infinite-dimensional statement or add the finite-dimensional qualifier to the abstract and introduction.","section":"Abstract; Theorem 4"},{"comment":"The proof contains two unsubstantiated steps. First, the claim that a completely intersubjective measurement on an n-dimensional system has at most n outcomes does not follow from the displayed linear-independence argument: the affine space of effects on an n-dimensional state space has dimension n+1, and the argument as written gives at most n+1 unless an unstated dimension convention is used; moreover, if zero effects are allowed, as they are in Example 2, one can pad a measurement with arbitrarily many zero outcomes. Second, the Krein–Millman step asserts that 1_S − εΣ a_i decomposes as an exact finite sum Σ b_j of indecomposable effects each of which is itself an effect; this is not automatic from Krein–Millman on a cone and requires an argument. Since Lemma 7 is presented as one of the two constructions behind Theorem 4, these gaps must be repaired or the lemma must be removed from","section":"End Matter, Lemma 7"},{"comment":"The key assertion 'Since a and b are indecomposable, the three-outcome measurement A = (λa, μb, 1−λa−μb) is intersubjective' is not proved and is not valid as stated for arbitrary indecomposable effects and arbitrary feasible λ, μ. In particular, if a = b and λ = μ = 1/2, the measurement is not intersubjective because λa is a common lower bound for the first two effects. The argument must use distinctness of a and b and the maximality of λ+μ to rule out a nonzero common lower bound between λa and the residual effect (and similarly for μb). Please supply this argument explicitly; as written, the contrapositive proof is incomplete at exactly the point where the assumption about three-outcome intersubjective measurements is invoked.","section":"End Matter, Lemma 6"}],"minor_comments":[{"comment":"Zero-effect outcomes should be explicitly excluded or handled throughout. Lemma 7's outcome-count bound and statement (iii) — that every completely intersubjective measurement admits states it perfectly distinguishes — are false if zero effects are counted as outcomes, since arbitrary zero effects can be added to a PVM or to any completely intersubjective measurement. Please state that the claims concern nonzero effects.","section":"End Matter, Lemma 7; Advantage in information processing, item (iii)"},{"comment":"The step 'Summing this inequality over all choices of complements ... both sides add up to 1_S' is quite terse. A more explicit partition argument showing that the diagonal entries equal the coarse-grained effects would improve readability and remove any doubt about the equality claim.","section":"Supplemental Appendix S4, Lemma S.6"},{"comment":"The symbol 'supp' is used for effects without definition, and the SOT-continuity of the projection meet used in Theorem S.1 is stated without proof. Since these are infinite-dimensional technical points, a brief justification or reference would be helpful.","section":"Theorem 2; Theorem S.1"},{"comment":"The statement that the states are 'perfectly distinguished only by A' requires qualification: if zero-effect outcomes are permitted, padding A with zeros yields other measurements that also perfectly distinguish the same states. The intended uniqueness claim should be made precise.","section":"Example 4"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the paper contains a genuinely interesting operational characterization of PVMs (Theorem 3), and the finite-dimensional classicality statement (Theorem 4) is plausible and likely correct. The main problems are (i) the abstract and introduction overclaim the classicality result by omitting the finite-dimensional qualifier, and (ii) the End Matter proofs, especially Lemma 7, contain gaps that are load-bearing for Theorem 4 as written. These are repairable within the scope of the manuscript — the authors can either add the missing arguments, weaken the advertised claims, or both. I would not recommend rejection; a careful major revision could make the paper acceptable. I also suggest that the authors state plainly that the classicality characterization is imported from finite-dimensional GPT results [30] rather than derived here."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real result here is Theorem 3: a POVM is a PVM iff every coarse-graining is intersubjective. I checked the proof skeleton; it is correct and genuinely operational, not just algebraic. The new notion of complete intersubjectivity does the work, and Proposition 1 (intersubjectivity iff sharpness) is a clean equivalence that ties the concept to Gudder's existing framework.\n\nWhat is done well: Proposition 1 is proved correctly, Theorem 2 (intersubjective POVMs have disjoint supports) holds even in infinite dimension, and the quantitative α-intersubjectivity is a nice addition with closed forms for qubit, classical, and coin-toss measurements. The authors also resolve Ozawa's open converse by constructing intersubjective POVMs that are not PVMs, and they do so without circularity—the load-bearing inputs (Gudder's sharp effects, the simplex characterization, Ozawa's forward direction) are external results.\n\nThe soft spots are real but patchable. The abstract and introduction claim that 'a system is classical iff intersubjectivity is preserved under any coarse-graining' in full generality, but Theorem 4 is proved only for finite-dimensional systems. That is not a harmless qualifier: the proof uses finite-dimensionality at structurally essential points, including the imported simplex characterization, Carathéodory's theorem, and the bound on the number of outcomes. The advertised claim outruns the supplied argument. Lemma 7 also asserts a finite Krein–Milman decomposition without proving the coefficient bound, and Lemma 6 compresses a maximality step. A referee should ask for those details, but nothing here suggests the central theorems are wrong.\n\nThe paper is for GPT foundations and quantum measurement people. It deserves a serious referee; I would recommend conditional acceptance after the authors fix the dimension overclaim and fill in the two terse proof steps.","headline":"Solid GPT result giving an operational characterization of PVMs via complete intersubjectivity; the classicality theorem is finite-dimensional despite the abstract's unqualified claim, and two proof steps need tightening, but the core is sound and worth refereeing.","tokens_in":20147,"tokens_out":2265,"would_cite":true,"duration_ms":23720,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A measurement is a projective observable exactly when two observers always agree on its outcome at every resolution.","keywords":["intersubjectivity","complete intersubjectivity","projective measurements","POVMs","generalized probabilistic theories","sharpness","coarse-graining","classicality"],"falsifier":"Find a qubit POVM that passes the complete-intersubjectivity test but is not a PVM, or find a finite-dimensional non-classical system in which every intersubjective measurement is completely intersubjective; either would refute the paper's two central theorems.","tokens_in":19066,"feed_emoji":"👥","tokens_out":7910,"duration_ms":74359,"temperature":0.7,"pith_summary":"This paper sets out to find an operational principle that singles out projective measurements (PVMs) — the traditional 'physical observables' — from the wider class of positive-operator-valued measures (POVMs). It claims that the principle is complete intersubjectivity: a measurement is a PVM exactly when two observers performing it simultaneously are guaranteed to obtain the same outcome and this guarantee survives every coarse-graining of the outcomes. It further claims that, in finite-dimensional general probabilistic theories, a system is classical if and only if every intersubjective measurement on it remains intersubjective under coarse-graining. If the claims hold, physical observables acquire a purely operational definition, and non-classicality is characterized by the way objective agreement can be lost when measurement resolution is reduced. The paper also quantifies the principle, computing the guaranteed agreement probability for coin-tossing, classical, and qubit measurements, and shows that intersubjective measurements suffice for state tomography and optimal state discrimination.","feed_headline":"Agreement at every resolution selects projective measurements","feed_subtitle":"A simple observer-agreement rule singles out projective measurements and flags non-classicality.","key_machinery":"The central object is complete intersubjectivity: a measurement for which the measurement and every coarse-graining are intersubjective, where a measurement is intersubjective when any joint implementation with itself gives identical outcomes to two simultaneous observers (equivalently, its only self-joint measurement is the 'agree always' one). The workhorse is the equivalence between intersubjectivity and sharpness: a measurement is sharp if no nonzero effect can be simultaneously bounded above by two distinct outcome effects; sharp effects are the generalization of projection operators. The proof structure also uses indecomposable effects (the extremal rays of the effect cone) and the str","core_discovery":"On its own terms, the paper's discovery is a pair of characterizations. In quantum theory, a POVM is a PVM if and only if it is completely intersubjective: every coarse-graining (merged outcome setting) is intersubjective, meaning the only joint measurement of the measurement with itself is the one in which both observers always read the same outcome. The proof uses an equivalence the authors establish between intersubjectivity and sharpness of effects — a measurement is sharp when no nonzero effect lies below two distinct outcome effects — and the fact that sharp effects in quantum theory are projections. In finite-dimensional general probabilistic theories, a system is classical if and onl","pith_inferences":["If complete intersubjectivity is taken as the definition of an observable in arbitrary probabilistic theories, it may become a tool for deriving or constraining theories from an 'observer agreement' axiom, possibly linking to axiomatic reconstructions of quantum theory.","The resolution-dependent loss of agreement resembles contextuality; a resource-theoretic formulation in which complete intersubjectivity measures 'sharpness resources' seems a natural next step, though the paper does not pursue it.","An empirical test: implement a qubit POVM known to be intersubjective but not projective, coarse-grain two of its outcomes, and check whether two detectors disagree more often than the projective bound; the paper's formulas predict where the divergence appears.","The classicality theorem is proved only for finite-dimensional systems; extending it to infinite-dimensional state spaces, where the decomposition arguments require topological care, is an open problem the paper leaves implicit."],"forward_implications":["In quantum theory, complete intersubjectivity becomes an operational definition of a physical observable, matching the algebraic definition of a projection-valued measure.","In any non-classical theory, there exist measurements whose outcome agreement is guaranteed only at full resolution; merging outcomes can destroy the observers' consensus.","Classical theories are exactly those in which intersubjectivity is never destroyed by coarse-graining, giving a new operational test for classicality.","The α-intersubjectivity value supplies a quantitatively interpretable 'observable-likeness' or 'noiselessness' measure for measurements, with simple closed-form expressions.","Intersubjective measurements are abundant enough to support state tomography and single-shot state discrimination in every general probabilistic theory."],"fun_headline_variants":["Observer agreement at every resolution singles out projective measurements","Agreement at every coarse-graining marks quantum observables","Intersubjectivity at all resolutions defines classicality","Agreement under all coarse-graining marks non-classicality"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The classicality characterization depends on cited structural facts about sharp effects and about decomposing the constant effect into indecomposable effects in a particular exact way, and if those facts fail the theorem's conclusion collapses.","fun_headline_variants_meta":{"raw":{"variants":["Observer agreement at every resolution singles out projective measurements","Agreement at every coarse-graining marks quantum observables","Intersubjectivity at all resolutions defines classicality","Agreement under all coarse-graining marks non-classicality"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000771,"raw_usage":{"total_tokens":3214,"prompt_tokens":673,"completion_tokens":2541,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":417,"completion_tokens_details":{"reasoning_tokens":2476}},"tokens_in":417,"tokens_out":2541,"duration_ms":19872,"temperature":1.0,"reasoning_tokens":2476,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T19:36:20.181451+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a qubit POVM that passes the complete-intersubjectivity test but is not a PVM, or find a finite-dimensional non-classical system in which every intersubjective measurement is completely intersubjective; either would refute the paper's two central theorems.","supporting_citations":[],"review_version":1}